[Soekris] Net5501: USB devices & attach/detach events

[Guy F. Boyd]

> From: Martin Johnson [mailto:martin.johnson.uk.lists at googlemail.com]
> Sent: Wednesday, May 28, 2008 10:19 AM
> To: soekris.support.list at vta.com
> Subject: Re: [Soekris] Net5501: USB devices & attach/detach events
>
>
> Didn't FreeBSD 6.x have a daemon called usbd, that needed to
> be run to
> get hotplugging to happen?    This daemon is no longer offered on
> FreeBSD 7.0, so I think the USB support there has been reworked.
>
> Fairly sure the 5501 has  USB 2.0, certainly it seems to run
> my Lacie
> drive at a decent speed under FreeBSD 7.0.
>
> - Martin
>
>

Hmm,  hadn't noticed that the 5501 actually is listed for 2.0. I guess I was
remembering the 4801 sheet from the website. You're absolutly right. ( I also
forgot about usbdevs(8)).

Well, the plot thickens a bit then.

With respect to usbd(8) and RELENG_6, usbd(8) is actually running here... but I
believe devd(8) gets at all of this via /var/run/devd.pipe whether usbd is
running or not.

Under a baseline test with other PC hardware ( and with any of my pile of USB
host interface cards ) the kernel interface reports the attach/detach events as
they happen, with or without any user space deamons running. On the 5501's
native host interface port, the previously described behavior is observed.

This is not a show-stopper for me but it is a curiousity. Seems to be consistent
between 2 5501s I have here right now.
Probably time to turn on some of those loud debug knobs in sys/dev/usb and see
what happens. I may also try RELENG_7 or openBSD just to see if there's any
difference.

Sure seems vaguely power or signel level related on the surface though, given
that the PCI host interface cards work...

Guy Boyd
Atlanta Ga USA